AI Music Brings The Sound Of Scammers To Spotify
(Bloomberg Opinion) -- “Mahogany Flex,” “Afternoon Nap” and “Tropical Taste” sound like bad cocktail ideas that never made it past the napkin stage. But they’re just a selection of the musical styles offered by artificial-intelligence app Boomy, which conjures up ready-made tracks for users in a matter of seconds— and which has become the latest target of record labels’ fight against “fake” music.
The industry’s issue with Boomy — which claims to have helped create 14.6 million songs, or 14% of all recorded music — is somewhat different to the sneaky deepfakes directly ripping off the likes of Drake and the Weeknd with chilling accuracy.
Spotify Technology SA has taken down a chunk of the app’s catalog from its streaming service after Universal Music Group NV raised suspicions that the songs’ popularity was being artificially inflated with fake streams, the Financial Times reported. This is the head-spinning flipside of the AI coin: Machines listening to music that machines helped create — and making real money from a $30 billion recorded music market.
If that’s what’s been going on — Boomy itself says it’s “categorically against” manipulated streams — there’s a clear need for action. One aim should be to build a better digital mousetrap to catch those skimming profits from the royalty pot. But longer term, record labels and streaming platforms need to fix the economic structure that siphons money away from artists lower down the musical food chain in the first place.
Fake streams are a big problem that’s been around for longer than Boomy, even if AI threatens to super-charge the potential for fraud. A study published by French music body CNM in January estimated 1% to 3% of French music streams in 2021 — as many as 3 billion — were fake. It identified several elaborate strategies: Click farms, repetitive re-uploading of tracks and fake playlists set to play on repeat are a few examples. They echo similar fakery in the online ad market and social media.
While streaming platforms ban these practices — as Spotify’s terms and conditions can attest — their economic incentives are to attract as much content as possible to reel in listeners, and their technological defenses aren’t invulnerable. In one recent legal tussle with an indie label, Spotify said it had “ inadvertently” paid out tens of thousands of dollars in royalty payments for streams that it said had been artificially inflated. Some kind of industry-wide clean-up is well overdue.
Yet it’s misleading to imagine this is a problem that will simply be solved by better fraud detection. The all-you-can-eat model built by record labels, rights holders and tech platforms has been its own worst enemy in some ways. The current system of distribution of royalties means listeners’ $9.99 subscription fees go into a big pot that is paid out according to streaming market share. When all of the spoils go to the most popular tracks that are listened to for more than 30 seconds, everyone’s habits have become more robotic: Songs have become shorter, playlists are stuffed with fictional artists churning out muzak, and real musicians are more prone to toying with half-minute formats, from songwriter Mac DeMarco’s recent 199-song nine-hour opus to French rapper Lorenzo’s 68-track album. No wonder it’s getting harder to spot the bots: There is more music on Apple Inc.’s streaming service than you could listen to in a lifetime.
Which is why one of the best ways to deter the bots might be to change the music payment model itself, as suggested by one indie-music representative at a recent UK parliamentary inquiry into the economics of streaming. A “user-centric” system that would see listeners’ subscription fees go to the artists they individually streamed would make it easier to stop money flowing to music that only machines want to hear. If the fear and hype around AI can bring balance to an unequal system, then maybe humans have a chance after all.
More From Bloomberg Opinion:
- What the Music Industry Must Do About AI: H. Drew Blackburn
- ChatGPT's Knock Knock Knockin' on Spotify's Door: Lionel Laurent
- Creative AI Is Generating Some Messy Problems: Parmy Olson
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Lionel Laurent is a Bloomberg Opinion columnist covering digital currencies, the European Union and France. Previously, he was a reporter for Reuters and Forbes.
More stories like this are available on bloomberg.com/opinion
©2023 Bloomberg L.P.